library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

Eigendata<-read.table(file = "~/GitHub/vespa_analyses/Input/Eigendata.txt", header=TRUE, sep="\t")
Eigendata_shortind<-read.table(file = "~/GitHub/vespa_analyses/Input/Eigendata_shortind.txt", header=TRUE, sep="\t")

library(ggplot2)
library(ggpubr)

Volgende datasets zijn gecreëerd:

Eigendata: dataset met alle vliegtijden en eigen metingen van de weersparameters
Eigendata_shortind: dataset met kortste vliegtijden per individu en eigen metingen van de weersparameters

Voor het model: Flight time ~ Distance gebruiken we de dataset kortste vliegtijden per individu Omdat we hiermee de theoretische regel 1min=100m kunnen verifiëren. De imkers nemen hiervoor altijd kortste meting

Voor modellen met weerparameters, gewicht, urbanisatie: Hiervoor nemen we telkens de hele dataset, omdat elke meting van deze factoren afhangt.

Outlier van nest 29, Bait 1, Individu A weggelaten in deze dataset. (600m op 2min lijkt wel heel snel)

Outlier van Melle ook weggelaten (meting op 2 km niet betrouwbaar)

Meting in Excel script wel nog terug te vinden.

1.1) ForagingSpeed vs Temperature

Graphs

Van temperatuur ook categorische variabele maken (aanraden van Prof.Vangestel)

Eigendata$Temperature_cat<-cut(Eigendata$Temperature, c(8,13,17,21,25,29,33))
ggplot(data=subset(Eigendata, !is.na(Temperature_cat)), aes(x=Temperature_cat)) + geom_bar(fill="lightcoral")

Model Output

Temperature = continuous variable

Niet significant
Normality niet ok!

Temperature = discrete variable

Een aantal klassen significant, interpretatie?
Normality niet ok!

Poisson verdeling ook geprobeerd met glmer maar foutmelding (zie email)

library(lmerTest)

## Loading required package: lme4

## Loading required package: Matrix

## 
## Attaching package: 'lmerTest'

## The following object is masked from 'package:lme4':
## 
##     lmer

## The following object is masked from 'package:stats':
## 
##     step

library(lme4)

model_temp_cont<-lmer(Flighttime_min ~ Temperature + (1|Individualcode), na.action=na.omit, data=Eigendata, offset= Distance)
summary(model_temp_cont)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Flighttime_min ~ Temperature + (1 | Individualcode)
##    Data: Eigendata
##  Offset: Distance
## 
## REML criterion at convergence: 1366.3
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.9375 -0.2898 -0.0230  0.1500  3.9428 
## 
## Random effects:
##  Groups         Name        Variance  Std.Dev.
##  Individualcode (Intercept) 67952.248 260.677 
##  Residual                       3.702   1.924 
## Number of obs: 183, groups:  Individualcode, 58
## 
## Fixed effects:
##               Estimate Std. Error         df t value Pr(>|t|)    
## (Intercept) -332.03927   34.45121   58.47880  -9.638  1.1e-13 ***
## Temperature    0.09884    0.19540  124.22687   0.506    0.614    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr)
## Temperature -0.113

anova(model_temp_cont, ddf="Satterthwaite", type=3)

## Type III Analysis of Variance Table with Satterthwaite's method
##              Sum Sq Mean Sq NumDF  DenDF F value Pr(>F)
## Temperature 0.94747 0.94747     1 124.23  0.2559 0.6138

res<-residuals(model_temp_cont)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.80863, p-value = 3.188e-14

qqnorm(res)
qqline(res)

model_temp_cat<-lmer(Flighttime_min ~ Temperature_cat + (1|Individualcode), offset=Distance, data=Eigendata)
summary(model_temp_cat)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Flighttime_min ~ Temperature_cat + (1 | Individualcode)
##    Data: Eigendata
##  Offset: Distance
## 
## REML criterion at convergence: 1329.5
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.7833 -0.2906 -0.0097  0.1433  3.7080 
## 
## Random effects:
##  Groups         Name        Variance Std.Dev.
##  Individualcode (Intercept) 68858.22 262.409 
##  Residual                       3.37   1.836 
## Number of obs: 183, groups:  Individualcode, 58
## 
## Fixed effects:
##                         Estimate Std. Error        df t value Pr(>|t|)    
## (Intercept)            -341.5757    37.1284   56.0945  -9.200 8.45e-13 ***
## Temperature_cat(13,17]    3.2036     0.8981  121.0057   3.567 0.000518 ***
## Temperature_cat(17,21]    4.7465     1.4043  121.0385   3.380 0.000976 ***
## Temperature_cat(21,25]    3.3026     2.3108  121.1265   1.429 0.155523    
## Temperature_cat(25,29]   63.4493    99.9309   56.0040   0.635 0.528058    
## Temperature_cat(29,33]    3.4708     2.7746  121.1038   1.251 0.213371    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) T_(13, T_(17, T_(21, T_(25,
## Tmp_(13,17] -0.021                            
## Tmp_(17,21] -0.026  0.639                     
## Tmp_(21,25] -0.028  0.389  0.608              
## Tmp_(25,29] -0.372  0.008  0.010  0.010       
## Tmp_(29,33] -0.025  0.324  0.506  0.833  0.009

anova(model_temp_cat, ddf="Satterthwaite", type=3)

## Type III Analysis of Variance Table with Satterthwaite's method
##                 Sum Sq Mean Sq NumDF  DenDF F value  Pr(>F)  
## Temperature_cat 53.121  10.624     5 97.929  3.1524 0.01109 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

res<-residuals(model_temp_cat)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.84698, p-value = 1.395e-12

qqnorm(res)
qqline(res)

1.2) Flight time vs Distance + Temperature

Graph

library(knitr)
library(kableExtra)

## Warning in !is.null(rmarkdown::metadata$output) && rmarkdown::metadata$output
## %in% : 'length(x) = 2 > 1' in coercion to 'logical(1)'

## 
## Attaching package: 'kableExtra'

## The following object is masked from 'package:dplyr':
## 
##     group_rows

library("plot3D")

x <- Eigendata$Distance
y <- Eigendata$Temperature
z <- Eigendata$Flighttime_min

fit <- lm(z ~ x + y, na.action=na.exclude)
x.pred <- seq(min(x[!is.na(x)]), max(x[!is.na(x)]), length.out = 20)
y.pred <- seq(min(y[!is.na(y)]), max(y[!is.na(y)]), length.out = 20)
xy <- expand.grid( x = x.pred, y = y.pred)
z.pred <- matrix(predict(fit, newdata = xy), 
                 nrow = 20, ncol = 20)
fitpoints <- predict(fit)

scatter3D(x, y, z, pch = 19, cex = 0.6, colvar=FALSE, col="dodgerblue3", theta = 210, phi = 10, bty="u", col.panel ="grey93", expand =0.4, col.grid = "white", xlab = "Distance", ylab = "Temperature", zlab = "Flight time", surf = list(x = x.pred, y = y.pred, z = z.pred,  
facets = TRUE, col=ramp.col(col = c("dodgerblue4", "seagreen2"), n = 100, alpha=0.8), fit = fitpoints, border="black"),main = "Flight time vs Distance + Temperature")

Met plotly

Had met deze link problemen met expand.grid https://stackoverflow.com/questions/38331198/add-regression-plane-to-3d-scatter-plot-in-plotly

Dan maar grid dat van hierboven gebruikt, maar nu klopt er precies iets niet? Alle datapunten liggen boven het oppervlak.

library(plotly)

## 
## Attaching package: 'plotly'

## The following object is masked from 'package:ggplot2':
## 
##     last_plot

## The following object is masked from 'package:stats':
## 
##     filter

## The following object is masked from 'package:graphics':
## 
##     layout

library(tidyverse)

## ── Attaching packages
## ───────────────────────────────────────
## tidyverse 1.3.2 ──

## ✔ tibble  3.1.8     ✔ purrr   1.0.1
## ✔ tidyr   1.2.1     ✔ stringr 1.5.0
## ✔ readr   2.1.4     ✔ forcats 1.0.0
## ── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
## ✖ tidyr::expand()          masks Matrix::expand()
## ✖ plotly::filter()         masks dplyr::filter(), stats::filter()
## ✖ kableExtra::group_rows() masks dplyr::group_rows()
## ✖ dplyr::lag()             masks stats::lag()
## ✖ tidyr::pack()            masks Matrix::pack()
## ✖ tidyr::unpack()          masks Matrix::unpack()

FlightsNoNA<-Eigendata %>%  filter(Flighttime_min!="NA")
FlightsNoNA<-FlightsNoNA %>%  filter(Temperature!="NA")
fig <- plot_ly(FlightsNoNA, x = ~Distance, y = ~Temperature, z = ~Flighttime_min, size=1)
fig <- fig %>% add_markers()
fig <- fig %>% layout(scene = list(xaxis = list(title = 'Distance'),
                     yaxis = list(title = 'Temperature'),
                     zaxis = list(title = 'Flight time (min)')))
p2 <- add_trace(p = fig,
                z = z.pred,
                x = seq(2100, 0, by = -100),
                y = seq(0, 30, by = 10),
                type = "surface")
p2

Model Ouput

Beide significant
Normality niet ok!

model_tempdist_all<-lmer(Flighttime_min ~ Distance + Temperature + (1|Individualcode), na.action=na.exclude, data=Eigendata)
summary(model_tempdist_all)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Flighttime_min ~ Distance + Temperature + (1 | Individualcode)
##    Data: Eigendata
## 
## REML criterion at convergence: 803.9
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -1.8894 -0.3936 -0.1599  0.1706  4.6284 
## 
## Random effects:
##  Groups         Name        Variance Std.Dev.
##  Individualcode (Intercept) 1.445    1.202   
##  Residual                   3.452    1.858   
## Number of obs: 183, groups:  Individualcode, 58
## 
## Fixed effects:
##              Estimate Std. Error        df t value Pr(>|t|)    
## (Intercept)  4.561776   0.845837 62.971937   5.393 1.11e-06 ***
## Distance     0.006563   0.000976 78.841216   6.725 2.52e-09 ***
## Temperature -0.147567   0.045732 80.427866  -3.227  0.00181 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) Distnc
## Distance     0.050       
## Temperature -0.909 -0.379

res<-residuals(model_tempdist_all)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.76816, p-value = 1.014e-15

qqnorm(res)
qqline(res)

2) ForagingSpeed vs Cloudcoverage

Graphs

ggplot(Eigendata, aes(x=Cloudcoverage, y=ForagingSpeed)) + geom_point(color="lightblue") + geom_smooth(method="lm", formula =y ~ x, se=TRUE, fullrange=FALSE, level=0.95, color="darkblue") + ggtitle("All data | eigen metingen")+ theme(plot.title = element_text(hjust = 0.5, size=10))

Model Ouput

Niet significant
Normality ok!

model_cloud_all<-lmer(Flighttime_min ~ Cloudcoverage +  (1|Individualcode), offset=Distance, na.action=na.omit, data=Eigendata)
summary(model_cloud_all)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Flighttime_min ~ Cloudcoverage + (1 | Individualcode)
##    Data: Eigendata
##  Offset: Distance
## 
## REML criterion at convergence: 1055.6
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.8981 -0.3081 -0.0411  0.1862  3.6991 
## 
## Random effects:
##  Groups         Name        Variance  Std.Dev.
##  Individualcode (Intercept) 81926.936 286.229 
##  Residual                       3.905   1.976 
## Number of obs: 152, groups:  Individualcode, 39
## 
## Fixed effects:
##               Estimate Std. Error       df t value Pr(>|t|)    
## (Intercept)   -315.994     45.836   38.007  -6.894 3.42e-08 ***
## Cloudcoverage   -1.897      1.090  112.017  -1.741   0.0844 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr)
## Cloudcoverg -0.010

anova(model_cloud_all, ddf="Satterthwaite", type=3)

## Type III Analysis of Variance Table with Satterthwaite's method
##               Sum Sq Mean Sq NumDF  DenDF F value  Pr(>F)  
## Cloudcoverage  11.84   11.84     1 112.02   3.032 0.08438 .
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

res<-residuals(model_cloud_all)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.8356, p-value = 9.028e-12

qqnorm(res)
qqline(res)

3.1) ForagingSpeed vs Windspeed

Graphs

Telkens voor de modellen

ForagingSpeed ~ Windspeed

ForagingSpeed ~ Windspeed²

plot1<-ggplot(Eigendata, aes(x=Windspeed, y=ForagingSpeed)) + geom_point(color="khaki3") + geom_smooth(method="lm", formula =y ~ x, se=TRUE, fullrange=FALSE, level=0.95, color="olivedrab") + ggtitle("All data | eigen metingen")+ theme(plot.title = element_text(hjust = 0.5, size=10))

plot2<-ggplot(Eigendata, aes(x=Windspeed, y=ForagingSpeed)) + geom_point(color="khaki3") + geom_smooth(method="lm", formula =y ~ I(x^2), se=TRUE, fullrange=FALSE, level=0.95, color="olivedrab") + ggtitle("All data | eigen metingen")+ theme(plot.title = element_text(hjust = 0.5, size=10))


ggarrange(plot1, plot2 + rremove("x.text"), 
          labels = c("A", "B"),
          ncol = 2, nrow = 1)

Model Output

Beide modellen licht significant
Beide normality niet ok!

model_wind_all<-lmer(Flighttime_min ~ Windspeed + (1|Individualcode), offset=Distance, na.action=na.omit, data=Eigendata)
summary(model_wind_all)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Flighttime_min ~ Windspeed + (1 | Individualcode)
##    Data: Eigendata
##  Offset: Distance
## 
## REML criterion at convergence: 1262.7
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.0307 -0.3235 -0.0332  0.1735  3.4539 
## 
## Random effects:
##  Groups         Name        Variance  Std.Dev.
##  Individualcode (Intercept) 68307.191 261.36  
##  Residual                       3.571   1.89  
## Number of obs: 174, groups:  Individualcode, 52
## 
## Fixed effects:
##              Estimate Std. Error        df t value Pr(>|t|)    
## (Intercept) -327.7082    36.2538   51.0530  -9.039 3.56e-12 ***
## Windspeed      0.9518     0.3951  121.0224   2.409   0.0175 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##           (Intr)
## Windspeed -0.023

anova(model_wind_all, ddf="Satterthwaite", type=3)

## Type III Analysis of Variance Table with Satterthwaite's method
##           Sum Sq Mean Sq NumDF  DenDF F value  Pr(>F)  
## Windspeed 20.731  20.731     1 121.02   5.805 0.01749 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

model_wind2_all<-lmer(Flighttime_min ~ I(Windspeed^2) + (1|Individualcode), offset=Distance, na.action=na.omit, data=Eigendata)
summary(model_wind2_all)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Flighttime_min ~ I(Windspeed^2) + (1 | Individualcode)
##    Data: Eigendata
##  Offset: Distance
## 
## REML criterion at convergence: 1264.8
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.0393 -0.3245 -0.0275  0.1740  3.4495 
## 
## Random effects:
##  Groups         Name        Variance  Std.Dev.
##  Individualcode (Intercept) 68276.457 261.298 
##  Residual                       3.551   1.884 
## Number of obs: 174, groups:  Individualcode, 52
## 
## Fixed effects:
##                  Estimate Std. Error         df t value Pr(>|t|)    
## (Intercept)    -326.99309   36.23952   51.01860  -9.023 3.79e-12 ***
## I(Windspeed^2)    0.24383    0.09524  121.02600   2.560   0.0117 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr)
## I(Wndspd^2) -0.014

anova(model_wind2_all, ddf="Satterthwaite", type=3)

## Type III Analysis of Variance Table with Satterthwaite's method
##                Sum Sq Mean Sq NumDF  DenDF F value Pr(>F)  
## I(Windspeed^2) 23.272  23.272     1 121.03   6.554 0.0117 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

res<-residuals(model_wind_all)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.83203, p-value = 7.048e-13

qqnorm(res)
qqline(res)

res<-residuals(model_wind2_all)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.83542, p-value = 9.856e-13

qqnorm(res)
qqline(res)

3.2) ForagingSpeed vs Windspeed | afhankelijk van de windrichting

Hiervoor werd telkens voor elke meting nagegaan of de hoornaar met meewind (tailwind), tegenwind (upwind) of loodrechte wind (perpendicular) te maken had. Dit volgens de formules:

|𝜃flight −𝜃wind| ≤ 45 is tailwind

45 < |𝜃flight −𝜃wind|<135 is (quasi) perpendicular

|𝜃 flight −𝜃wind| ≥135 upwind

Windrichting werd telkens bepaald t.o.v. de vliegrichting (Wind_flight) en t.o.v. richting tot het nest (Wind_nest).

Graphs

ggplot(data=subset(Eigendata, !is.na(Wind_flight)), aes(x= Wind_flight, col=Wind_flight, y=ForagingSpeed)) + geom_boxplot()

## Warning: Removed 6 rows containing non-finite values (`stat_boxplot()`).

ggplot(data=subset(Eigendata, !is.na(Wind_nest)), aes(x= Wind_nest, col=Wind_nest, y=ForagingSpeed)) + geom_boxplot()

## Warning: Removed 6 rows containing non-finite values (`stat_boxplot()`).

ggplot(data=subset(Eigendata, !is.na(Wind_flight)), aes(x= Windspeed, col=Wind_flight, y=ForagingSpeed)) + geom_point() + facet_grid(~Wind_flight) + geom_smooth(method="lm", formula = y ~ I(x^2), se=TRUE, fullrange=FALSE, level=0.95)

## Warning: Removed 6 rows containing non-finite values (`stat_smooth()`).

## Warning: Removed 6 rows containing missing values (`geom_point()`).

ggplot(data=subset(Eigendata, !is.na(Wind_nest)), aes(x= Windspeed, col=Wind_nest, y=ForagingSpeed)) + geom_point() + facet_grid(~Wind_nest) + geom_smooth(method="lm", formula = y ~ I(x^2), se=TRUE, fullrange=FALSE, level=0.95)

## Warning: Removed 6 rows containing non-finite values (`stat_smooth()`).
## Removed 6 rows containing missing values (`geom_point()`).

Model Output

Anova van Foragingspeed en de 3 windinvloeden

Dit voor windrichting t.o.v. richting tot het nest en t.o.v. vliegrichting
Niet significant
Normality niet oke
interpretatie?

windflightanova<-lmer(Flighttime_min ~ Wind_flight + (1|Individualcode), na.action=na.exclude, offset=Distance, data=Eigendata)
summary(windflightanova)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Flighttime_min ~ Wind_flight + (1 | Individualcode)
##    Data: Eigendata
##  Offset: Distance
## 
## REML criterion at convergence: 1151.3
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.8767 -0.3156 -0.0390  0.1724  3.8245 
## 
## Random effects:
##  Groups         Name        Variance  Std.Dev.
##  Individualcode (Intercept) 65409.102 255.752 
##  Residual                       3.965   1.991 
## Number of obs: 161, groups:  Individualcode, 46
## 
## Fixed effects:
##                      Estimate Std. Error        df t value Pr(>|t|)    
## (Intercept)         -319.3628    37.7103   45.0036  -8.469 7.23e-11 ***
## Wind_flighttailwind    0.1535     1.3936  113.0331   0.110    0.913    
## Wind_flightupwind     -0.8702     0.8884  113.0167  -0.980    0.329    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) Wnd_flghtt
## Wnd_flghttl -0.006           
## Wnd_flghtpw -0.006  0.089

res<-residuals(windflightanova)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.82875, p-value = 1.873e-12

qqnorm(res)
qqline(res)

windnestanova<-lmer(Flighttime_min ~ Wind_nest + (1|Individualcode), na.action=na.exclude, offset=Distance, data=Eigendata)
summary(windnestanova)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Flighttime_min ~ Wind_nest + (1 | Individualcode)
##    Data: Eigendata
##  Offset: Distance
## 
## REML criterion at convergence: 1149.6
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.8920 -0.3231 -0.0462  0.1811  3.8447 
## 
## Random effects:
##  Groups         Name        Variance  Std.Dev.
##  Individualcode (Intercept) 65436.126 255.805 
##  Residual                       3.923   1.981 
## Number of obs: 161, groups:  Individualcode, 46
## 
## Fixed effects:
##                    Estimate Std. Error        df t value Pr(>|t|)    
## (Intercept)       -319.1543    37.7196   45.0110  -8.461 7.41e-11 ***
## Wind_nesttailwind    0.4132     1.4495  113.0518   0.285    0.776    
## Wind_nestupwind     -1.6444     1.1318  113.0384  -1.453    0.149    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) Wnd_nstt
## Wnd_nsttlwn -0.008         
## Wnd_nstpwnd -0.008  0.000

res<-residuals(windnestanova)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.83112, p-value = 2.345e-12

qqnorm(res)
qqline(res)

ForagingSpeed vs WindSpeed voor Tailwind dataset

Significant
Normality niet ok!

data_tailwind<-subset(Eigendata, Wind_flight == "tailwind")

model_tailwind<-lmer(Flighttime_min ~ Windspeed + (1|Individualcode), offset= Distance, na.action=na.omit, data=data_tailwind)
summary(model_tailwind)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Flighttime_min ~ Windspeed + (1 | Individualcode)
##    Data: data_tailwind
##  Offset: Distance
## 
## REML criterion at convergence: 158.2
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -1.74181 -0.18655 -0.03152  0.01763  1.88834 
## 
## Random effects:
##  Groups         Name        Variance Std.Dev.
##  Individualcode (Intercept) 46780.21 216.287 
##  Residual                      10.78   3.283 
## Number of obs: 21, groups:  Individualcode, 8
## 
## Fixed effects:
##              Estimate Std. Error        df t value Pr(>|t|)   
## (Intercept) -1109.162    240.549     5.999  -4.611  0.00365 **
## Windspeed     314.022    104.860     5.999   2.995  0.02418 * 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##           (Intr)
## Windspeed -0.948

anova(model_tailwind, ddf="Satterthwaite", type=3)

## Type III Analysis of Variance Table with Satterthwaite's method
##           Sum Sq Mean Sq NumDF  DenDF F value  Pr(>F)  
## Windspeed 96.645  96.645     1 5.9991   8.968 0.02418 *
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

res<-residuals(model_tailwind)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.82908, p-value = 0.001892

qqnorm(res)
qqline(res)

ForagingSpeed vs WindSpeed voor Perpendicular dataset

Significant
Normality niet ok!

data_perpendicular<-subset(Eigendata, Wind_flight == "perpendicular")

model_perpendicular<-lmer(Flighttime_min ~ Windspeed + (1|Individualcode), offset= Distance, na.action=na.omit, data=data_perpendicular)

## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
## unable to evaluate scaled gradient

## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
## Model failed to converge: degenerate Hessian with 1 negative eigenvalues

summary(model_perpendicular)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Flighttime_min ~ Windspeed + (1 | Individualcode)
##    Data: data_perpendicular
##  Offset: Distance
## 
## REML criterion at convergence: 730.6
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.0291 -0.3957 -0.0109  0.2275  3.5200 
## 
## Random effects:
##  Groups         Name        Variance  Std.Dev.
##  Individualcode (Intercept) 55413.701 235.401 
##  Residual                       2.319   1.523 
## Number of obs: 104, groups:  Individualcode, 33
## 
## Fixed effects:
##              Estimate Std. Error        df t value Pr(>|t|)    
## (Intercept) -307.2799    40.9876   32.0278  -7.497 1.55e-08 ***
## Windspeed      1.9726     0.4035   70.0147   4.888 6.23e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##           (Intr)
## Windspeed -0.021
## optimizer (nloptwrap) convergence code: 0 (OK)
## unable to evaluate scaled gradient
## Model failed to converge: degenerate  Hessian with 1 negative eigenvalues

res<-residuals(model_perpendicular)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.88169, p-value = 1.377e-07

qqnorm(res)
qqline(res)

ForagingSpeed vs WindSpeed voor Upwind dataset

Niet significant
Normality niet ok!

data_upwind<-subset(Eigendata, Wind_flight == "upwind")

model_upwind<-lmer(Flighttime_min ~ Windspeed + (1|Individualcode), offset=Distance, na.action=na.omit, data=data_upwind)
summary(model_upwind)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Flighttime_min ~ Windspeed + (1 | Individualcode)
##    Data: data_upwind
##  Offset: Distance
## 
## REML criterion at convergence: 248.3
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -1.65651 -0.32560 -0.01373  0.16570  2.43454 
## 
## Random effects:
##  Groups         Name        Variance  Std.Dev.
##  Individualcode (Intercept) 38919.718 197.281 
##  Residual                       2.771   1.665 
## Number of obs: 36, groups:  Individualcode, 12
## 
## Fixed effects:
##             Estimate Std. Error       df t value Pr(>|t|)   
## (Intercept) -201.008     57.112   11.125  -3.520  0.00472 **
## Windspeed     -1.029      2.015   23.052  -0.511  0.61442   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##           (Intr)
## Windspeed -0.075

res<-residuals(model_upwind)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.8614, p-value = 0.0003521

qqnorm(res)
qqline(res)

4) ForagingSpeed vs Weight individual

Graphs

ggplot(Eigendata, aes(x=Weight_ind, y=ForagingSpeed)) + geom_point(color="plum") + geom_smooth(method="lm", formula =y ~ x, se=TRUE, fullrange=FALSE, level=0.95, color="darkorchid4") + ggtitle("All data")+ theme(plot.title = element_text(hjust = 0.5, size=10))

Model Output

Niet significant
Normality ok!

model_weight_all<-lmer(Flighttime_min ~ Weight_ind + (1|Individualcode), offset=Distance, na.action=na.omit, data=Eigendata)
summary(model_weight_all)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Flighttime_min ~ Weight_ind + (1 | Individualcode)
##    Data: Eigendata
##  Offset: Distance
## 
## REML criterion at convergence: 557.1
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.2045 -0.3196 -0.0862  0.1576  3.4639 
## 
## Random effects:
##  Groups         Name        Variance  Std.Dev.
##  Individualcode (Intercept) 79697.832 282.308 
##  Residual                       3.194   1.787 
## Number of obs: 89, groups:  Individualcode, 20
## 
## Fixed effects:
##             Estimate Std. Error      df t value Pr(>|t|)
## (Intercept)  -301.93     341.39   18.00  -0.884    0.388
## Weight_ind     39.82    1000.02   18.00   0.040    0.969
## 
## Correlation of Fixed Effects:
##            (Intr)
## Weight_ind -0.983

anova(model_weight_all, ddf="Satterthwaite", type=3)

## Type III Analysis of Variance Table with Satterthwaite's method
##               Sum Sq   Mean Sq NumDF  DenDF F value Pr(>F)
## Weight_ind 0.0050637 0.0050637     1 17.999  0.0016 0.9687

res<-residuals(model_weight_all)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.77318, p-value = 2.123e-10

qqnorm(res)
qqline(res)

5.1) ForagingSpeed vs Urbanisation (Traject 100m)

Graphs

Urbanisatiecirkels niet berekend, meteen trajecten genomen want die zijn betere maatstaf

Verder werken met 100m buffer!

# Meting in urbaan gebied met koude temperatuur weglaten (Nest 38 en 39, 3 wiekpotten)

Eigendata2<-Eigendata %>%  filter(Traject100m!=0.504960)
Eigendata2<-Eigendata2 %>%  filter(Traject100m!=0.496950)
Eigendata2<-Eigendata2 %>%  filter(Traject100m!=0.411898)


ggplot(data=subset(Eigendata, !is.na(Observer)), aes(x=Traject100m, y=ForagingSpeed)) + geom_point(col="slateblue1") + geom_abline(intercept=3.8984, slope= -5.3143, color="slateblue4", linewidth=1) + ggtitle("Met outlier | random effects") + xlim(0, 0.6)

## Warning: Removed 26 rows containing missing values (`geom_point()`).

ggplot(data=subset(Eigendata2, !is.na(Observer)), aes(x=Traject100m, y=ForagingSpeed)) + geom_point(col="slateblue1") + geom_abline(intercept=3.7838, slope= -4.0061, color="slateblue4", linewidth=1) + ggtitle("Zonder outlier | random effects") + xlim(0, 0.6)

## Warning: Removed 26 rows containing missing values (`geom_point()`).

Verder werken met 100m buffer!

Graph Temperature vs Urbanisation

ggplot(data=subset(Eigendata, !is.na(Temperature_cat)), aes(x=Temperature_cat, col=Temperature_cat, y=Traject100m)) + geom_boxplot()

Model Output

Met outlier

Hoogsignificant
Normality niet ok!

model_urb1<-lmer(Flighttime_min ~ Traject100m + (1|Individualcode), offset=Distance, na.action=na.omit, data=Eigendata)
summary(model_urb1)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Flighttime_min ~ Traject100m + (1 | Individualcode)
##    Data: Eigendata
##  Offset: Distance
## 
## REML criterion at convergence: 1918
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.0862 -0.2381 -0.0067  0.1117  4.1024 
## 
## Random effects:
##  Groups         Name        Variance  Std.Dev.
##  Individualcode (Intercept) 69478.334 263.587 
##  Residual                       3.445   1.856 
## Number of obs: 234, groups:  Individualcode, 94
## 
## Fixed effects:
##             Estimate Std. Error      df t value Pr(>|t|)    
## (Intercept)  -551.80      42.05   92.00 -13.124  < 2e-16 ***
## Traject100m  1018.13     225.44   92.00   4.516 1.86e-05 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr)
## Traject100m -0.763

anova(model_urb1, ddf="Satterthwaite", type=3)

## Type III Analysis of Variance Table with Satterthwaite's method
##             Sum Sq Mean Sq NumDF  DenDF F value    Pr(>F)    
## Traject100m 70.262  70.262     1 91.998  20.397 1.863e-05 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

res<-residuals(model_urb1)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.77701, p-value < 2.2e-16

qqnorm(res)
qqline(res)

Zonder outlier

Significant
Normality niet ok!

model_urb2<-lmer(Flighttime_min ~ Traject100m + (1|Individualcode), offset=Distance, na.action=na.omit, data=Eigendata2)
summary(model_urb2)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Flighttime_min ~ Traject100m + (1 | Individualcode)
##    Data: Eigendata2
##  Offset: Distance
## 
## REML criterion at convergence: 1679.8
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.7342 -0.2214 -0.0047  0.0920  4.0368 
## 
## Random effects:
##  Groups         Name        Variance  Std.Dev.
##  Individualcode (Intercept) 73469.487 271.053 
##  Residual                       2.352   1.534 
## Number of obs: 199, groups:  Individualcode, 89
## 
## Fixed effects:
##             Estimate Std. Error      df t value Pr(>|t|)    
## (Intercept)  -550.98      46.42   87.00 -11.870  < 2e-16 ***
## Traject100m  1009.41     291.35   87.00   3.465 0.000827 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr)
## Traject100m -0.785

anova(model_urb2, ddf="Satterthwaite", type=3)

## Type III Analysis of Variance Table with Satterthwaite's method
##             Sum Sq Mean Sq NumDF  DenDF F value    Pr(>F)    
## Traject100m 28.229  28.229     1 86.999  12.004 0.0008265 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

res<-residuals(model_urb2)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.71442, p-value < 2.2e-16

qqnorm(res)
qqline(res)

5.2) ForagingSpeed vs Urbanisation (Categorisch)

ggplot(data=Eigendata, aes(x=Urbanisation, y=ForagingSpeed, col=Urbanisation)) + geom_boxplot()

## Warning: Removed 26 rows containing non-finite values (`stat_boxplot()`).

model_urb_cat<-lmer(Flighttime_min ~ Urbanisation + (1|Individualcode), offset=Distance, na.action=na.omit, data=Eigendata )
summary(model_urb_cat)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: Flighttime_min ~ Urbanisation + (1 | Individualcode)
##    Data: Eigendata
##  Offset: Distance
## 
## REML criterion at convergence: 1906.2
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.0847 -0.2381 -0.0062  0.1113  4.1025 
## 
## Random effects:
##  Groups         Name        Variance  Std.Dev.
##  Individualcode (Intercept) 67030.070 258.902 
##  Residual                       3.445   1.856 
## Number of obs: 234, groups:  Individualcode, 94
## 
## Fixed effects:
##                           Estimate Std. Error      df t value Pr(>|t|)    
## (Intercept)                -524.37      40.94   91.00 -12.809  < 2e-16 ***
## UrbanisationRandstedelijk   106.72      61.41   91.00   1.738   0.0856 .  
## UrbanisationVerstedelijkt   346.48      68.72   91.00   5.042 2.34e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) UrbnsR
## UrbnstnRnds -0.667       
## UrbnstnVrst -0.596  0.397

6) Flight error vs urbanisation (25m, 50m, 100m)

Flight error is the difference between the actual angle from the bait to the nest and the flight direction I measured.

Graphs

plot28<-ggplot(data=subset(Eigendata, !is.na(Observer)), aes(x=Traject25m, y=FlightError)) + geom_point(col="hotpink1") + geom_smooth(method="lm", formula =y ~ x, se=TRUE, fullrange=FALSE, level=0.95, color="violetred4")

plot29<-ggplot(data=subset(Eigendata, !is.na(Observer)), aes(x=Traject50m, y=FlightError)) + geom_point(col="hotpink1") + geom_smooth(method="lm", formula =y ~ x , se=TRUE, fullrange=FALSE, level=0.95, color="violetred4")

plot30<-ggplot(data=subset(Eigendata, !is.na(Observer)), aes(x=Traject100m, y=FlightError)) + geom_point(col="hotpink1")+ geom_smooth(method="lm", formula =y ~ x, se=TRUE, fullrange=FALSE, level=0.95, color="violetred4")

ggarrange(plot28, plot29, plot30 + rremove("x.text"), 
          labels = c("A", "B", "C"),
          ncol = 2, nrow = 2)

## Warning: Removed 9 rows containing non-finite values (`stat_smooth()`).

## Warning: Removed 9 rows containing missing values (`geom_point()`).

## Warning: Removed 9 rows containing non-finite values (`stat_smooth()`).

## Warning: Removed 9 rows containing missing values (`geom_point()`).

## Warning: Removed 9 rows containing non-finite values (`stat_smooth()`).

## Warning: Removed 9 rows containing missing values (`geom_point()`).

Model Output

Allemaal niet significant
Allemaal normality niet ok!

model_errurb25<-lmer(FlightError ~ Traject25m + (1|NestID) + (1|NestID:BaitID) + (1|Individualcode), na.action=na.omit, data=Eigendata)
summary(model_errurb25)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: FlightError ~ Traject25m + (1 | NestID) + (1 | NestID:BaitID) +  
##     (1 | Individualcode)
##    Data: Eigendata
## 
## REML criterion at convergence: 1947.1
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.7734 -0.1939 -0.0500  0.0181  6.9743 
## 
## Random effects:
##  Groups         Name        Variance Std.Dev.
##  Individualcode (Intercept) 133.89   11.571  
##  NestID:BaitID  (Intercept) 119.73   10.942  
##  NestID         (Intercept)  14.35    3.788  
##  Residual                    55.40    7.443  
## Number of obs: 251, groups:  Individualcode, 111; NestID:BaitID, 85; NestID, 34
## 
## Fixed effects:
##             Estimate Std. Error     df t value Pr(>|t|)    
## (Intercept)   17.796      2.744 39.401   6.484 1.05e-07 ***
## Traject25m    -3.487     14.699 48.243  -0.237    0.813    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##            (Intr)
## Traject25m -0.704

anova(model_errurb25, ddf="Satterthwaite", type=3)

## Type III Analysis of Variance Table with Satterthwaite's method
##            Sum Sq Mean Sq NumDF  DenDF F value Pr(>F)
## Traject25m  3.119   3.119     1 48.243  0.0563 0.8135

model_errurb50<-lmer(FlightError ~ Traject50m + (1|NestID) + (1|NestID:BaitID) + (1|Individualcode), na.action=na.omit, data=Eigendata)
summary(model_errurb50)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: FlightError ~ Traject50m + (1 | NestID) + (1 | NestID:BaitID) +  
##     (1 | Individualcode)
##    Data: Eigendata
## 
## REML criterion at convergence: 1946.9
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.7832 -0.1962 -0.0514  0.0206  6.9719 
## 
## Random effects:
##  Groups         Name        Variance Std.Dev.
##  Individualcode (Intercept) 134.66   11.604  
##  NestID:BaitID  (Intercept) 120.80   10.991  
##  NestID         (Intercept)  12.35    3.514  
##  Residual                    55.33    7.438  
## Number of obs: 251, groups:  Individualcode, 111; NestID:BaitID, 85; NestID, 34
## 
## Fixed effects:
##             Estimate Std. Error     df t value Pr(>|t|)    
## (Intercept)   16.620      2.846 40.187   5.841  7.8e-07 ***
## Traject50m     5.634     15.381 43.317   0.366    0.716    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##            (Intr)
## Traject50m -0.734

anova(model_errurb50, ddf="Satterthwaite", type=3)

## Type III Analysis of Variance Table with Satterthwaite's method
##            Sum Sq Mean Sq NumDF  DenDF F value Pr(>F)
## Traject50m 7.4251  7.4251     1 43.317  0.1342 0.7159

model_errurb100<-lmer(FlightError ~ Traject100m + (1|NestID) + (1|NestID:BaitID) + (1|Individualcode), na.action=na.omit, data=Eigendata)
summary(model_errurb100)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: FlightError ~ Traject100m + (1 | NestID) + (1 | NestID:BaitID) +  
##     (1 | Individualcode)
##    Data: Eigendata
## 
## REML criterion at convergence: 1946.8
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -3.7833 -0.1966 -0.0513  0.0225  6.9715 
## 
## Random effects:
##  Groups         Name        Variance Std.Dev.
##  Individualcode (Intercept) 134.40   11.593  
##  NestID:BaitID  (Intercept) 121.48   11.022  
##  NestID         (Intercept)  11.92    3.452  
##  Residual                    55.33    7.439  
## Number of obs: 251, groups:  Individualcode, 111; NestID:BaitID, 85; NestID, 34
## 
## Fixed effects:
##             Estimate Std. Error     df t value Pr(>|t|)    
## (Intercept)   16.575      2.962 40.888   5.597 1.63e-06 ***
## Traject100m    5.851     16.003 41.504   0.366    0.717    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr)
## Traject100m -0.759

anova(model_errurb100, ddf="Satterthwaite", type=3)

## Type III Analysis of Variance Table with Satterthwaite's method
##             Sum Sq Mean Sq NumDF  DenDF F value Pr(>F)
## Traject100m 7.3957  7.3957     1 41.504  0.1337 0.7165

res<-residuals(model_errurb25)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.58149, p-value < 2.2e-16

qqnorm(res)
qqline(res)

res<-residuals(model_errurb50)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.58086, p-value < 2.2e-16

qqnorm(res)
qqline(res)

res<-residuals(model_errurb100)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.581, p-value < 2.2e-16

qqnorm(res)
qqline(res)

Model Selection

fullmodel<-lmer(ForagingSpeed ~ Traject100m + Windspeed + Temperature + Cloudcoverage + Weight_ind + (1|NestID) + (1|NestID:BaitID) + (1|Individualcode), na.action=na.omit, data=Eigendata)

## boundary (singular) fit: see help('isSingular')

summary(fullmodel)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: 
## ForagingSpeed ~ Traject100m + Windspeed + Temperature + Cloudcoverage +  
##     Weight_ind + (1 | NestID) + (1 | NestID:BaitID) + (1 | Individualcode)
##    Data: Eigendata
## 
## REML criterion at convergence: 170
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.7356 -0.5550  0.1066  0.5685  1.8733 
## 
## Random effects:
##  Groups         Name        Variance Std.Dev.
##  Individualcode (Intercept) 0.03398  0.1843  
##  NestID:BaitID  (Intercept) 0.31924  0.5650  
##  NestID         (Intercept) 0.00000  0.0000  
##  Residual                   0.30817  0.5551  
## Number of obs: 88, groups:  Individualcode, 20; NestID:BaitID, 15; NestID, 10
## 
## Fixed effects:
##               Estimate Std. Error       df t value Pr(>|t|)   
## (Intercept)    4.99457    2.20523 29.00627   2.265  0.03117 * 
## Traject100m   -7.86760    2.58297 18.69255  -3.046  0.00674 **
## Windspeed     -0.17727    0.16151  9.98177  -1.098  0.29815   
## Temperature   -0.03233    0.08430 27.81931  -0.384  0.70425   
## Cloudcoverage  1.31882    0.39406 24.27275   3.347  0.00266 **
## Weight_ind    -1.44193    2.55293 10.61380  -0.565  0.58394   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) Trj100 Wndspd Tmprtr Cldcvr
## Traject100m -0.789                            
## Windspeed    0.204 -0.409                     
## Temperature -0.888  0.775 -0.384              
## Cloudcoverg  0.241 -0.388  0.127 -0.327       
## Weight_ind  -0.545  0.188  0.084  0.150 -0.048
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')

model1<-lmer(ForagingSpeed ~ Traject100m + Windspeed + Cloudcoverage + Weight_ind + (1|NestID) + (1|NestID:BaitID) + (1|Individualcode), na.action=na.omit, data=Eigendata)

## boundary (singular) fit: see help('isSingular')

summary(model1)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: 
## ForagingSpeed ~ Traject100m + Windspeed + Cloudcoverage + Weight_ind +  
##     (1 | NestID) + (1 | NestID:BaitID) + (1 | Individualcode)
##    Data: Eigendata
## 
## REML criterion at convergence: 167
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.6747 -0.5380  0.1197  0.5758  1.9362 
## 
## Random effects:
##  Groups         Name        Variance Std.Dev.
##  Individualcode (Intercept) 0.03317  0.1821  
##  NestID:BaitID  (Intercept) 0.28865  0.5373  
##  NestID         (Intercept) 0.00000  0.0000  
##  Residual                   0.30791  0.5549  
## Number of obs: 88, groups:  Individualcode, 20; NestID:BaitID, 15; NestID, 10
## 
## Fixed effects:
##               Estimate Std. Error      df t value Pr(>|t|)   
## (Intercept)     4.2462     0.9816 10.7873   4.326  0.00126 **
## Traject100m    -7.0951     1.5703  8.0472  -4.518  0.00193 **
## Windspeed      -0.1992     0.1439  9.4887  -1.384  0.19790   
## Cloudcoverage   1.2675     0.3639 22.4685   3.483  0.00206 **
## Weight_ind     -1.3126     2.4408 10.0199  -0.538  0.60249   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) Trj100 Wndspd Cldcvr
## Traject100m -0.343                     
## Windspeed   -0.325 -0.195              
## Cloudcoverg -0.121 -0.231  0.017       
## Weight_ind  -0.906  0.115  0.155  0.007
## optimizer (nloptwrap) convergence code: 0 (OK)
## boundary (singular) fit: see help('isSingular')

model2<-lmer(ForagingSpeed ~ Traject100m + Windspeed + Cloudcoverage + (1|NestID) + (1|NestID:BaitID) + (1|Individualcode), na.action=na.omit, data=Eigendata)
summary(model2)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: ForagingSpeed ~ Traject100m + Windspeed + Cloudcoverage + (1 |  
##     NestID) + (1 | NestID:BaitID) + (1 | Individualcode)
##    Data: Eigendata
## 
## REML criterion at convergence: 343.7
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -2.46705 -0.53111  0.00685  0.48244  2.85559 
## 
## Random effects:
##  Groups         Name        Variance Std.Dev.
##  Individualcode (Intercept) 0.09978  0.3159  
##  NestID:BaitID  (Intercept) 0.39334  0.6272  
##  NestID         (Intercept) 0.25303  0.5030  
##  Residual                   0.37363  0.6113  
## Number of obs: 152, groups:  Individualcode, 39; NestID:BaitID, 26; NestID, 15
## 
## Fixed effects:
##                Estimate Std. Error        df t value Pr(>|t|)    
## (Intercept)     4.99137    0.43247  22.10904  11.542 7.87e-11 ***
## Traject100m    -8.09188    1.46500  13.92745  -5.523 7.64e-05 ***
## Windspeed      -0.29289    0.09917  81.93301  -2.953   0.0041 ** 
## Cloudcoverage   0.32661    0.27303 124.69507   1.196   0.2339    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) Trj100 Wndspd
## Traject100m -0.661              
## Windspeed   -0.454 -0.049       
## Cloudcoverg -0.180 -0.230  0.106

model3<-lmer(ForagingSpeed ~ Traject100m + Windspeed + (1|NestID) + (1|NestID:BaitID) + (1|Individualcode), na.action=na.omit, data=Eigendata)
summary(model3)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: ForagingSpeed ~ Traject100m + Windspeed + (1 | NestID) + (1 |  
##     NestID:BaitID) + (1 | Individualcode)
##    Data: Eigendata
## 
## REML criterion at convergence: 412.1
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -2.53652 -0.49591  0.01821  0.49551  2.83727 
## 
## Random effects:
##  Groups         Name        Variance Std.Dev.
##  Individualcode (Intercept) 0.08236  0.2870  
##  NestID:BaitID  (Intercept) 0.93614  0.9675  
##  NestID         (Intercept) 0.02917  0.1708  
##  Residual                   0.37294  0.6107  
## Number of obs: 174, groups:  Individualcode, 52; NestID:BaitID, 37; NestID, 20
## 
## Fixed effects:
##              Estimate Std. Error        df t value Pr(>|t|)    
## (Intercept)   4.87080    0.36651  30.78189  13.290 2.72e-14 ***
## Traject100m  -7.39979    1.32907  19.45673  -5.568 2.09e-05 ***
## Windspeed    -0.20583    0.09498 138.14425  -2.167    0.032 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) Trj100
## Traject100m -0.655       
## Windspeed   -0.551 -0.018

res<-residuals(model3)
shapiro.test(res)

## 
##  Shapiro-Wilk normality test
## 
## data:  res
## W = 0.98787, p-value = 0.1401

qqnorm(res)
qqline(res)

library(cAIC4)

## Loading required package: stats4

## Loading required package: nlme

## 
## Attaching package: 'nlme'

## The following object is masked from 'package:lme4':
## 
##     lmList

## The following object is masked from 'package:dplyr':
## 
##     collapse

library(MASS)

## 
## Attaching package: 'MASS'

## The following object is masked from 'package:plotly':
## 
##     select

## The following object is masked from 'package:dplyr':
## 
##     select

library(AICcmodavg)

## 
## Attaching package: 'AICcmodavg'

## The following object is masked from 'package:lme4':
## 
##     checkConv

AIC(fullmodel)

## [1] 190.0063

AIC(model1)

## [1] 185.0204

AIC(model2)

## [1] 359.7193

AIC(model3)

## [1] 426.0821

Multicollinearity check

library(ggcorrplot)
library(car)

## Loading required package: carData

## 
## Attaching package: 'car'

## The following object is masked from 'package:purrr':
## 
##     some

## The following object is masked from 'package:dplyr':
## 
##     recode

datacor<-Eigendata[, c("Temperature", "Cloudcoverage", "Windspeed", "Weight_ind", "Traject100m", "NestID")]
source("~/GitHub/vespa_analyses/Input/HighstatLibV10.R") 
corvif(datacor)

## 
## 
## Variance inflation factors
## 
##                   GVIF
## Temperature   6.692872
## Cloudcoverage 1.356257
## Windspeed     1.332447
## Weight_ind    1.530327
## Traject100m   6.778816
## NestID        2.226214

cormat <- round(cor(datacor, use = "pairwise.complete.obs"), 2)
ggcorrplot(cormat, lab= TRUE, type = "lower", ggtheme = ggplot2::theme_gray,
   colors = c("#6D9EC1", "white", "#E46726"))

plot(Temperature~Traject100m, xlab="Urbanisation", data=Eigendata)

Eerder toeval dat Temperature en Traject100m gecorreleeerd zijn?

Eigendata$NestID<-as.factor(Eigendata$NestID)
ggplot(Eigendata, aes(x=NestID, col=NestID, y=Temperature)) + geom_boxplot()

## Warning: Removed 68 rows containing non-finite values (`stat_boxplot()`).

Masterthesis script, eigen weerparameters

Heleen Van Ransbeeck

2023-01-03

1.1) ForagingSpeed vs Temperature

Graphs

Model Output

1.2) Flight time vs Distance + Temperature

Graph

Model Ouput

2) ForagingSpeed vs Cloudcoverage

Graphs

Model Ouput

3.1) ForagingSpeed vs Windspeed

Graphs

Model Output

3.2) ForagingSpeed vs Windspeed | afhankelijk van de windrichting

Graphs

Model Output

Anova van Foragingspeed en de 3 windinvloeden

ForagingSpeed vs WindSpeed voor Tailwind dataset

ForagingSpeed vs WindSpeed voor Perpendicular dataset

ForagingSpeed vs WindSpeed voor Upwind dataset

4) ForagingSpeed vs Weight individual

Graphs

Model Output

5.1) ForagingSpeed vs Urbanisation (Traject 100m)

Graphs

Graph Temperature vs Urbanisation

Model Output

Met outlier

Zonder outlier

5.2) ForagingSpeed vs Urbanisation (Categorisch)

6) Flight error vs urbanisation (25m, 50m, 100m)

Graphs

Model Output

Model Selection

Multicollinearity check